3U8 KAF1L) POPULATION GROWTH-II
differentials in migration that will not vary, to some degree at least, in pattern and intensity with time and place. (Italics in original.)
Possibly age selectivity is the one exception to this statement, but otherwise it is incumbent upon any investigation of migrant selectivity at least to consider the possibility of change in migratory pattern over time.
Studies of migrant selectivity that depend entirely on census information are inadequate. Censuses do provide valuable data, and census bureaus should be encouraged to publish more of the information on migration which they have available, but the census enumeration instrument simply is too restricted in the amount and kind of questions it can cover.
Sample surveys are more flexible and can provide a much wider range of information. The Monterrey Mobility Study convinced me that reliable life histories can be obtained and that the computer makes feasible for the first time the processing and analysis of large numbers of cases. (See 55.) Life histories lend themselves very well to cohort analysis, and this is indispensable for the survey analyst who wishes to escape from static analysis.
Sample surveys on migration are now appearing with increasing frequency, but like the Monterrey study, they are often handicapped in making migrant-native comparisons because of a restriction to one locale. This limitation can be overcome through the use of large national surveys; that by Blau and Duncan (43) is an excellent example. But in order to permit adequate analysis by size of community and by region, a large sample number (N) is required, perhaps as much as 10,000. Such surveys are expensive and difficult to execute in most developing countries, but they provide the best way of determining the importance of various features of migration, including selectivity, on the national level. And if they could be replicated after 10 years or so, one would be able to say with some confidence whether patterns of migrant selectivity are changing.
Finally, the recent surge in output of studies dealing with migration makes it propitious to begin the arduous and tedious task of codifying the findings of these studies, both on a regional and world basis. This task, and my paper is only a gesture in that direction, not only would serve to clarify what we now know about migration, it would tell us where the major gaps in our knowledge are and how they could best be filled.
Notwithstanding the limitations that we have indicated, certain conclusions can be reached on the basis of our survey. Migrants to large cities in developing countries are heterogeneous in their socioeconomic backgrounds. This observation may seem so self-evident as to be trivial, but its acceptance allows us to clear away some of the misconceptions and errors that have occurred frequently in the literature. A number of social scientists have seen the migration process through a glass all-too-darkly. I have tried to explain the group who must make the (rreatest adjustment in stvlp. of livino nnH tVip.rpfnrp nrp nrp«nmali1\/ m^et inhe predominantly male rural-urban migration, in the multilingual character of the society, in the existence of polygamy, and